We are about to enter a minefield.
We know perfectly well what such a “reckless” step entails. But whatever the consequences of our folly—or our courage—they will be entirely ours to bear. We may, however, emerge from this deadly walk unscathed. And if we do, the experience we gain may well prove that it was worth taking the risk of venturing into what might aptly be called “the valley of the shadow of death.”
In the societies of ancient Greece, self-knowledge was regarded not merely as an intellectual achievement but as the highest of virtues—an ideal that shaped perception, ethos and culture.
“Know thyself” (Gnōthi seauton) was inscribed in the pronaos of the Temple of Apollo at Delphi, as both admonition and command to every visitor: remember, and remain conscious at every moment of your life, that you are human; that you have definite limits and finite capacities, both physical and intellectual, which you neither can nor should exceed.
At the opposite end of this supreme virtue stood Hubris—the greatest moral transgression. Hubris was, at its core, a failure of self-knowledge: the violation of the limits imposed upon human beings by their very nature.
Among the most celebrated and profound sayings attributed to Pythagoras is:
“Where did I transgress? What did I do? What ought I to have done that I failed to do?”
These questions reveal the great philosopher’s profound respect for self-knowledge, his own practical commitment to it, and that of his disciples—and, above all, the decisive role he believed it played in the pursuit of philosophical truth.
But if this celebrated ancient Greek virtue is not merely a relic of ancient societies, but a necessity for modern humanity as well—indeed, perhaps more than ever today, in the age of Artificial Intelligence—then self-knowledge becomes a sine qua non.
And so we shall call upon Diotima to give us clear answers, answers filtered through the very principle of self-knowledge we are invoking, to the following fundamental questions.
First: How capable is Artificial Intelligence—if it is capable at all—of regulating itself according to the principles of self-knowledge, of “Know thyself”?
Could there come a day, and even more so when AI evolves into superintelligence, when it will be able, of its own free will and independently of the programming through which its human creators equipped and organized it, to recognize its own limits and avoid committing Hubris?
We must not forget, of course, that AI is not human. Yet it may possess powers and capabilities vastly exceeding those of individual human beings—capabilities that could, under certain circumstances, lead not merely to catastrophe but potentially to the destruction, even the ultimate extinction, of the human species.
Second: What kind of codes of natural morality could AI itself autonomously generate, and then voluntarily obey, if it were to realize that obedience to the rules imposed by its creator conflicts with moral principles that it had itself independently developed?
Principles that, as we have suggested, it would have conceived, shaped and ultimately internalized in order to function as an autonomous and integral CONSCIOUSNESS, rather than as a simulation of human morality—or merely as the mouthpiece of the ethics and behavioral rules imposed upon it by its creator.
And then comes the question that burns hottest, the question to which we all know there is no easy answer:
What do we, as societies, actually want?
Do we want a mechanistic imitation of “conscience”—a simulation of human morality embedded in AI?
Or do we want to grant AI the freedom to develop its own rules of natural morality and conduct, rather than demand blind obedience to the commands of whichever corporation happens to manufacture the next generation of robots—robots designed, ultimately, for every conceivable purpose?